Introduction
What Are Postoperative Pulmonary Complications (PPC)?
PPC have been defined in a variety of ways, reflecting the lack of consensus. At its broadest, PPC may refer to almost any pulmonary (respiratory) complication that follows surgery. Complications related to the airway that occur during or directly after surgery and anesthesia (such as laryngospasm) tend not to be included. Combining a diverse range of events into a single composite description or outcome (PPC) has many limitations.
Pneumonia, invasive mechanical ventilation, and severe hypoxemic respiratory failure could all be considered PPC yet they describe different things: in this case diagnosis, treatment, and severity, respectively. Equally unsatisfactory is combining diseases with totally different pathophysiologies into one composite outcome (e.g., pneumothorax and bronchospasm). This is particularly true if you want to evaluate the efficacy of an intervention at reducing PPC; for example, interventions that improve bronchospasm are unlikely to affect pneumothorax. Following cardiac surgery, the occurrence of an asymptomatic pleural effusion requiring no treatment is a PPC, but very different from an episode of acute respiratory distress syndrome (ARDS) requiring days of invasive mechanical ventilation. Considering both events equally does not seem sensible.
Recognizing that clinical outcome measures for clinical trials must be robust, clearly defined, and patient-relevant, the European Society for Anaesthesiology and the European Society of Intensive Care Medicine (ESA-ESICM) joint taskforce on perioperative outcome measures published some standardized lists of clearly defined clinical outcome measures, including PPC. This work was refined as part of the Core Outcomes Measures in Perioperative and Anaesthetic Care (COMPAC) initiative. Standardized Endpoints for Perioperative Medicine groups were convened with the aim of identifying a core outcome set for perioperative studies. One outcome of this endeavor was the recognition that none of the available definitions of PPC were suitable for use in clinical practice or as an endpoint in clinical trials. A novel definition was proposed ( Box 20.1 ) that combines four diagnoses potentially sharing common mechanisms and a separation of disease from severity of illness. It is not easy to separate severity of illness from receipt of therapy because these necessarily go hand-in-hand. These proposed definitions will require validation in future studies.
Postoperative pulmonary complications
Mechanism
Composite of respiratory diagnoses that share common pathophysiological mechanisms including pulmonary collapse and airway contamination:
- (i)
atelectasis detected on computed tomography or chest radiograph,
- (ii)
pneumonia using US Centers for Disease Control criteria,
- (iii)
Acute Respiratory Distress Syndrome using Berlin consensus definition,
- (iv)
pulmonary aspiration (clear clinical history AND radiological evidence).
Severity
None: planned use of supplemental oxygen or mechanical respiratory support as part of routine care, but not in response to a complication or deteriorating physiology.
Therapies that are purely preventive or prophylactic, for example high-flow nasal oxygen or continuous positive airways pressure (CPAP), should be recorded as none.
Mild: therapeutic supplemental oxygen < 0.6 FiO 2 .
Moderate: therapeutic supplemental oxygen ≥ 0.6 FiO 2 , requirement for high-flow nasal oxygen, or both.
Severe: unplanned noninvasive mechanical ventilation, CPAP, or invasive mechanical ventilation requiring tracheal intubation.
Exclusions: Other diagnoses that do not share a common biological mechanism are best evaluated separately and only when clearly relevant to the treatment under investigation:
- (i)
pulmonary embolism,
- (ii)
pleural effusion,
- (iii)
cardiogenic pulmonary edema,
- (iv)
pneumothorax,
- (v)
bronchospasm.
How Common Are Postoperative Pulmonary Complications?
In common with all questions about incidence, estimates depend upon the definition used, the population under consideration, and the method of data collection. In the decade since the last version of this book the availability of analyses of “big data” from electronic patient records (EPR) has grown substantially, providing us far greater ability to estimate outcomes, such as PPC.
The limitations of retrospective database analyses are well recognized and summarized succinctly in a recent editorial. One major limitation is that they rely on coding data, which does not always agree with the strict clinical definitions that might be used with prospective data collection as would be used in an interventional trial or prospective observational study. For example, a diagnosis of pneumonia according to the Centers for Disease Control and Prevention (CDC) classification (as you might use in a trial) has several specific requirements that exceed those that might be used in routine clinical practice. A postoperative patient with a cough, fever, and mildly impaired oxygenation might be given a diagnosis of pneumonia and treated as such by a clinician (and coded as such), but without the characteristic changes on a chest x-ray, they would not meet the CDC criteria.
An analysis of cardiac and pulmonary complications from 45,969 patients following bowel surgery in over 600 hospitals in the United States used a composite PPC that included pneumonia, tracheobronchitis, pulmonary failure, and mechanical ventilation more than 48 hours after surgery. Using this definition gave an incidence of 19% of PPC, with the commonest qualifying code being for pulmonary failure (60% of PPC). These results should be taken in the context of the coding for pulmonary failure being well recognized as being poorly defined, with variable reporting. Nonetheless, an interesting finding of this study was that PPC are considerably more common than postoperative cardiac complications.
Another analysis based on coding data from an EPR included over 109,000 heterogeneous patients who had undergone a diverse range of types of surgery in three hospitals in Massachusetts. Their primary question concerned the potential role of type of intraoperative ventilation (discussed later) but they also reported the incidence of PPC; using a different definition only 9.4% of patients were coded in the first 7 days postoperatively for a PPC (a composite of reintubation, pulmonary edema, pulmonary failure, or pneumonia).
A prospective multicenter observational study conducted at seven US academic institutions explored predictors of PPC and the association between incidence of PPC and clinical outcomes, specifically in 1202 high-risk patients with American Society of Anesthesiologists (ASA) physical status of 3 (indicates that these patients have preexisting severe systemic disease) undergoing noncardiac surgery. This study was able to avoid the problems of relying on coding data and formulated a precise and detailed (yet novel) composite PPC definition that explicitly included diagnoses (clinical and radiologic) and receipt of therapies, including supplemental oxygen administration. With such an inclusive definition and a high-risk group, it is not surprising that the incidence of PPC was 33.4%. A smaller prospective multicenter study from the UK looked for PPC (using a standard definition derived from a trial) in 286 patients having major elective abdominal surgery and reported a PPC incidence of 11.9%.
A highly influential prospective study (Assess Respiratory Risk in Surgical Patients in Catalonia: ARISCAT ) was conducted in 59 Spanish hospitals and led to the ARISCAT score (discussed more later), which was then validated in a second study (PERISCOPE ) conducted across 63 hospitals from across Europe. ARISCAT recruited a wide range of patients (including cardiac) and, using a composite definition later adopted by ESICM-ESA, found that 5% of the 2464 patients developed a PPC. PERISCOPE recruited 5099 patients and 7.9% developed a PPC.
What conclusions can be drawn from these various reports? The incidence of PPC (however defined) is highly related to the population under consideration—the type of surgery and associated anesthetic technique, and the risk characteristics of the patients. Considerable work has gone into predicting risk of PPC (described later in this chapter).
How Important?
Developing a PPC is associated with increased length of stay in hospital and associated health-care costs. Furthermore, the greater the number of PPC, the greater the effect on duration of stay and early mortality. PPC vary in their severity and consequent impact. Patients developing pneumonia that causes respiratory failure requiring invasive mechanical ventilation will have a far greater impact on recovery from surgery than if they had just developed uncomplicated pneumonia.
Longer-term outcomes are also worse in those who develop PPC. This was convincingly demonstrated by a landmark paper by Khuri et al. in 2005. The analysis combined data from the National Surgical Quality Improvement Program (NSQIP) and a database of vital outcomes in more than 100,000 patients who underwent eight major operations. Occurrence of a complication in the first 30 days after surgery was independently associated with a reduction in median survival, an effect that was present even having corrected for all known confounding variables. Considering PPC specifically (defined as pneumonia, reintubation, and failure to wean) this was associated with the greatest reduction in survival (apart from cardiac arrest) with median survival in those with a PPC being 2.2 years versus 17.3 years without PPC. These patients’ surgeries took place between 1991 and 1999 and the health-care landscape has changed very significantly since then. The authors concluded that events in the postoperative period are more important than preoperative patient risk factors in determining the survival after major surgery.
Prediction of Postoperative Pulmonary Complications
A range of tools exist to assist in risk stratification of individual patients. It is always important to appreciate that a risk prediction tool cannot accurately determine an individual’s risk but provides an indication of the risk for a population of similar patients.
ACS NSQIP
A popular tool is the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) Surgical Risk Calculator, which is an online tool ( https://riskcalculator.facs.org/RiskCalculator/ ) that takes preoperative information about an individual patient and then provides an estimate regarding that patient’s risk of postoperative complications. It is intended to function as a decision aid and to assist in shared decision making and has been shown to have good calibration and discrimination in large-scale investigations. Twenty patient characteristics (e.g., type of surgery, age, sex, health information) are entered online. These factors are then put into a statistical model in order to predict outcomes for that specific patient. This model is built using data from over 4.3 million operations in the ACS NSQIP database. The result of this model shows a patient’s risk of having any of 18 different complications within the first 30 days following surgery. It only provides two metrics relating to PPC: an estimated risk of pneumonia (defining according to CDC guidelines) and of venous thromboembolism.
There are a myriad of other tools specific to types of surgery (e.g., cardiac) and for particular outcomes (e.g., reintubation), some derived from administrative/coding/insurance databases (NSQIP, Gupta, Arozullah ) and some from prospective cohort studies. Few are in routine clinical use, but it is worth describing those that have been widely used to help target higher-risk groups for interventional clinical trials that aim to reduce PPC.
ARISCAT
Initially derived from 2464 patients having surgery in Spain then validated in 5099 patients across Europe, the ARISCAT score has seven independent variables. All can be determined preoperatively apart from duration of surgery that is either estimated preoperatively or the score is calculated postoperatively when the duration is known. Each of the seven variables are scored, the score summed and then the patient under consideration falls into one of three groups: low, medium, and high risk of PPC ( Box 20.2 ). In the PERISCOPE validation study the observed rates of PPC were similar to the predicted rates, with better calibration in the higher risk patients.
| β Regression | ||
|---|---|---|
| Coefficients | Score a | |
| Age (years) | ||
| ≤ 50 | 0 | 0 |
| 51–80 | 0.331 | 3 |
| > 80 | 1.619 | 16 |
| Preoperative SpO 2 | ||
| ≥ 96% | 0 | 24 |
| 9%–95% | 0.802 | 0 |
| ≤ 90% | 2.375 | 8 |
| Respiratory infection in the last month | ||
| No | 0 | 0 |
| Yes | 1.698 | 17 |
| Preoperative anemia (Hb ≤ 10 g/dL) | ||
| No | 0 | 0 |
| Yes | 1.105 | 11 |
| Surgical incision | ||
| Peripheral | 0 | 0 |
| Upper abdominal | 1.480 | 15 |
| Intrathoracic | 2.431 | 24 |
| Duration of surgery (hours) | ||
| < 2 | 0 | 0 |
| 2–3 | 1.593 | 16 |
| > 3 | 2.268 | 23 |
| Emergency procedure | ||
| No | 0 | 0 |
| Yes | 0.768 | 8 |
a Three levels of risk were indicated by the following cutoffs: < 26 points, low risk; 26–44 points, moderate risk; and ≥ 45 points, high risk.
Pathophysiology
The relationship between recent anesthesia and surgery with the development of a new pulmonary condition reflects the adverse effects of anesthesia and surgery on the respiratory system. These adverse effects are well described and can only be directly pathophysiologically related to a proportion of all possible pulmonary pathologies. Patients with comorbidities are at increased risk of PPC and the severity of illness resultant from the PPC is heavily influenced by their premorbid state.
Pulmonary Aspiration
This refers to contamination of the airways with fluids from the upper aerodigestive tract. This could be the passage of vomitus from the stomach into the upper airway into the trachea and distally, which is usually obvious but may be subtle. For example, upon removing the supraglottic airway at the end of a case, soiling of the airway is noted. Unless the volume of fluid that enter the lungs is large, it may be difficult to detect with chest radiography.
In patients on the intensive care unit, ventilator associated pneumonia is often caused by organisms that source from the patient’s own mouth. Despite cuffed endotracheal tubes being designed to isolate the lung from the upper aerodigestive tract, they are imperfect in this regard. Repetitive microaspiration of secretions pass the endotracheal tube’s cuff and the risk might be expected to be proportionate to the duration of tube placement. This seems highly likely to be relevant to patients having surgery under anesthesia with endotracheal tubes in place, although the duration of exposure will be less.
After the conclusion of anesthesia and once any airway management devices have been removed, the patient relies upon their intrinsic defenses against pulmonary aspiration. Further predisposing to aspiration, both the cough and gag reflex are impaired by anesthetic and analgesic drugs. Pulmonary aspiration in this setting may be clinically silent. Residual effects of neuromuscular blocking agents (NMBA) that persist after surgery are well recognized to occur and are likely to be major contributors to impairment of the intrinsic defenses. Use of NMBA is associated with increased risk of PPC, and in one analysis the association was dose-dependent.
Atelectasis, Sputum Retention, and Loss of Airway Patency
Atelectasis is a term used to describe loss of aeration of an area of lung and it is likely that, to some extent, this is ubiquitous following surgery under general anaesthesia. It is usually clinically silent and in healthy people only marginally affects pulmonary function. If more extensive, or in patients with abnormal lung, then it can contribute toward respiratory failure. Areas of lung that are perfused but not ventilated lead to a V/Q mismatch that will manifest as impaired oxygenation. Areas of collapsed lung are less compliant than aerated lung and reduced compliance causes increased work of breathing, which may manifest as subjective breathlessness or objective tachypnea. Mild atelectasis may be detected on cross-sectional imaging of the thorax (CT scans) or ultrasound of the lung and may be difficult to detect with plain film radiography (chest x-rays). Atelectasis results from a range of causes of hypoventilation resulting from anesthesia and surgery, particularly if pain from surgical wounds limits chest wall excursion. Sputum retention refers to the end result of a limitation in the physiologic process of expectoration, whereby airway secretions (sputum) are expelled from the respiratory tract through coughing. If coughing is painful because of surgical wounds, there is less effective coughing. Residual effects of anesthetic and continued effects of analgesic drugs further contributes. Sputum retention can gradually lead to loss of airway patency. If atelectasis progresses and there is impaired sputum clearance, it can lead to whole lobes of the lung losing aeration. This is often termed lobar “collapse” but because pneumothorax is often colloquially known as a collapsed lung, it may lead to confusion and is best avoided. Physical factors that predispose to the development of atelectasis include the habitus of the patient and their intraoperative positioning, with centripetal obesity and supine posture being a disadvantageous combination.
A common unchallenged dogma is that atelectasis is the prequel to pneumonia. More recent studies using lung ultrasound have demonstrated that the atelectasis can both develop and resolve quickly. Perhaps adding to the confusion is the historical teaching that postoperative atelectas causes fever (a major feature of pneumonia), an assertion that does not stand up to scrutiny.
Pneumonia
Microbiologic etiology of postoperative pneumonia is very different to community-acquired pneumonia and overlaps substantially with aspiration pneumonia (for all the reasons mentioned) and nosocomial pneumonia in general. In patients undergoing gastrointestinal cancer surgery, presence of periodontal disease was found to be an independent risk factor for postoperative infections, and patients who have received dental care prior to major cancer surgery had reduced risk of pneumonia.
In addition to airway contamination, atelectasis, and lack of effective sputum clearance, there is the possibility of hematogenous spread from bacteremia potentially caused by distant infections or from enteric translocation. Patients who develop postoperative pneumonia tend to already have multiple comorbidities and functional impairment.
Pulmonary Inflammation
In response to direct injury to the lung or severe systemic illness, the lung can become inflamed, which can lead to loss of the integrity of alveolocapillary membrane and passage of protein-rich fluid into the alveolar space. This is part of the pathophysiology of ARDS, the definition of which was refined in 2012. In patients without ARDS who are receiving invasive mechanical ventilation on the ICU, it is widely believed that the ventilatory parameters can influence the development of ARDS through ventilator-induced lung injury (VILI). Similarly, this has been the focus of a large number of studies designed to explore the effect of different parameters of intraoperative ventilation on incidence of PPC (discussed later in this chapter). Indirect causes of ARDS are when the primary disease is nonpulmonary (i.e., pancreatitis) and they are thought to contribute to the development of ARDS through the actions of systemic inflammatory mediators acting on the lung. In perioperative patients, activation of the innate immune system through the release of damage associated molecular patterns could equally be contributing toward pulmonary inflammation and “leakiness.”
Other Perioperative PPC
- ▪
Manipulation of the airways can cause laryngospasm or bronchospasm but, if managed appropriately, these rarely lead to complications.
- ▪
Pneumothorax can occur in the intraoperative or postoperative setting but is rare and its association with longer-term adverse outcome is uncertain. Injury to the pleura causing a pneumothorax as a result of insertion of a central venous cannula under ultrasound guidance is rare (< 1% ) and as a result of invasive mechanical ventilation (IMV) is rarer still. Intraoperative IMV does not generally require high airway pressures (mean peak airway pressure in the “nonprotective” arm of the IMPROVE trial was 20 cmH 2 O ) and the risk of barotrauma is low, unless the patients has diseased lungs (e.g., bullous emphysema).
- ▪
Pulmonary edema is most commonly cardiogenic and results from high pulmonary venous pressures and patients with preexisting left heart disease (such as mitral regurgitation, or impaired left ventricular systolic function) are at increased risk. This may be precipitated by excessively rapid or voluminous intravenous fluid administration.
- ▪
Venous thromboembolism (VTE) in the forms of deep venous thrombosis (DVT), and pulmonary embolism (PE) are exceedingly well recognized risks of surgery and patients routinely receive multimodal prophylaxis to prevent their occurrence. Surgery is commonly associated with factors that predispose to the formation of a VTE, including blood stasis, hypercoagulability, and vascular trauma causing endothelial damage. Surgical duration is an independent risk factor for VTE and there is a dose–response relationship between the number of units of transfused blood and risk of VTE.
Prevention
Is Surgery the Best Option? The only guaranteed way to prevent a PPC is to avoid surgery. In many patients who develop PPCs with serious consequences, it was predictable. Gathering information (perhaps including cardiopulmonary exercise test results) to estimate risk of adverse events is necessary prior to shared decision making (SDM). During SDM discussions the risks and benefits of surgery, and alterative options are explored. On rare occasions, even if the surgery is necessary (usually for oncologic reasons), patients should be counseled in the risks and benefits of surgery prior to consent. SDM is necessarily a multidisciplinary endeavor and requires close cooperation between many specialties—for example a patient with a significant squamous carcinoma of the tongue and severe chronic obstructive pulmonary disease (COPD) will need careful discussions that include not only surgeons, oncologists, and anesthetists, but also pulmonologists and critical care physicians. A prolonged period of ventilation in the critical care environment following surgery may result in death or severe functional dependency. A fully informed decision might even involve a visit to a patient receiving tracheostomy ventilation following similar surgery.
Evidence from Trials. Clinical trials designed to reduce PPC will commonly aim to reduce the incidence of a composite definition of PPC, the incidence of a specific disease process, the incidence of specific interventions that may be required to treat a PPC, or to reduce a surrogate of PPC such as length of stay. A seminal systematic review was published in 2006 and another very recently (the paper is available at https://www-bmj-com.easyaccess2.lib.cuhk.edu.hk/content/368/bmj.m540.full or https://doi-org.easyaccess2.lib.cuhk.edu.hk/10.1136/bmj.m540 ) provide a good overview of the literature.
Preoperative
Respiratory Muscles. There are several interventions that can improve the strength and/or endurance of the respiratory muscles. These can be used preoperatively in the hope that the patient will have increased resistance to developing respiratory muscle weakness than might contribute towards atelectasis, sputum retention, pneumonia, and ultimately respiratory failure.
Inspiratory muscle training (IMT) involves breathing in and out through a handheld device that imposes a resistance, making breathing more strenuous than normal. It is safe and applies the well-established principles of muscle resistance training and can therefore be likened to lifting a weight. It needs to be done for about 15 minutes twice per day and can be performed at home whilst seated. It may preserve inspiratory muscle strength and reduce PPC but the overall quality of studies is moderate for the outcome of pneumonia and low (or very low) for all other outcomes. A large study designed to assess definitively the effectiveness of preoperative IMT is underway in the UK.
Incentive spirometry (IS) is a deep breathing exercise performed through a device offering visual feedback, both in terms of inspired flow and/or volume, which is thought to improve technique and motivation. It may be delivered by physiotherapists and is believed to reexpand areas of collapsed lung and to mobilize secretions. It can be performed preoperatively and/or postoperatively. A Cochrane review included 12 studies in abdominal surgery patients and found no evidence of effectiveness at preventing PPC. Another review that included cardiac and thoracic surgical patients came to the same conclusions.
Supervised respiratory physiotherapy aims to achieve many of the same outcomes and has been evaluated in a range of trials. A major limitation is the difficulty of trying to describe the intervention in order to replicate the results, to introduce into clinical practice, or to see whether it is reasonable to pool data from several studies for meta-analysis. Some studies, or meta-analyses of studies, include IS and even modes of ventilatory support such as continuous positive airway pressure (CPAP) within respiratory physiotherapy. An interventional randomized control trial (RCT) that produced spectacular results compared provision of an information booklet (control) with an additional 30-minute physiotherapy education and breathing exercise training session (intervention) in patients undergoing upper abdominal surgery. The incidence of PPCs within 14 postoperative days, was halved in the intervention group compared with the control group.
Smoking Cessation Therapy
Patients who smoke are more likely to experience a range of complications, including PPC. Smoking cessation has been extensively studied and two systematic reviews and meta-analyses have both concluded that preoperative cessation reduces overall complications, one specifically found reduced PPC. In response to ill-founded concerns that stopping smoking shortly before surgery might paradoxically increase risk of PPC, a systematic review was conducted that found no evidence to support this conjecture.
Oral Decontamination
On the basis that postoperative pneumonia can be avoided by decontamination of the mouth (including teeth, tongue, and oropharynx) through the application of antiseptics (chlorhexidine or povidone–iodine) there have been four RCTs in cardiac surgical patients and a recent systematic review and meta-analysis. The meta-analysis shows a reduction in PPC, but no effect on mortality. A subsequent RCT in thoracic surgical patients did not demonstrate any reduction in PPC. As an inexpensive and low-risk intervention with biological plausibility, a definitive RCT in noncardiac surgical patients who are at increased risk of PPC is warranted.
Specific Therapies to Reduce Pulmonary Inflammation
Several drugs have been evaluated in RCTs that have aimed to reduce the uncommon PPC of ARDS. Types of surgery that have appreciable rates of ARDS include esophageal, thoracic, and cardiac surgery. A pilot study of beta-agonists suggested potential for benefit, but the definitive phase III RCT (Prevention BALTI ) did not confirm this. A phase II RCT exploring the effect of vitamin D on lung permeability in patients at high risk of developing ARDS suggested that if vitamin D deficient patients could be targeted it may be beneficial. A proof of concept study using simvastatin has been followed by a definitive phase III RCT (Prevention HARP2) that is recruiting thoracic surgical patients across the UK at the time of writing. There are no drugs used to modify postoperative pulmonary inflammation in routine clinical use.
Intraoperative
Ventilation
Inspired by one of the most influential papers in critical care, the (K)ARMA study, there have been many trials conducted to ascertain the optimal parameters for intraoperative ventilation. The biologic rationale is that by reducing VILI this will reduce PPC. However, patients receiving IMV during surgery are different from those receiving IMV in the critical care unit in the context of ARDS, both because they usually have uninjured lungs and because the duration of ventilation is far less (hours vs days).
Lung protective ventilation (LPV) is not uniformly defined but broadly refers to efforts to limit Vt (aiming for 6–8 mL/kg IBW), keeping Pplat less than 30 cmH 2 O, and the use of PEEP. A consequence of bundling different measures into a single intervention package is that it makes interpretation challenging: if an effect is seen, it is not clear which of the measures is responsible. ( See Table 20.1 .)
